[Computer Vision 9] Structure-from-Motion Revisited

Review of Structure-from-Motion

Correspondence Search:

Find scene overlap in the input images, identify projection of the matched points between overlapping images.

Feature extraction --> Matching --> Geometric verification (e.g. RANSAC)

Incremental Reconstruction:

Initialization (choose the densest (has most overlapping cameras) pair of views to initialize the model) 

--> Image registration (solve the Perspective-n-Point problem (2D-3D features correspondences, to estimate pose P and intrinsic parameters of cameras) to add new images to the current model) / Triangulation (introduce new scene points into the set of points, increase stability of  existing model through redundancy) [Image registration and triangulation are two separate procedures, but their results are highly correlated]

--> Bundle adjustment

Scene Graph Augmentation

Multi-model geometric verification : select only non-panoramic and preferably calibrated image pairs, improve robustness of triangulation and subsequent image registrations.

Next Best View Selection

Track the number of visible points (see most triangulated points) and their distribution (uniform distribution, large covariance) in each candidate image, to minimize  the reconstruction error and maximize reconstruction robustness.

Robust and Efficient Triangulation

Formulate multi-view triangulation using RANSAC to get rid of outliers. 

Bundle Adjustment

Parameterization (choose robust loss function, BA) --> Filtering (after BA, we filter observations with large reprojection errors) --> Re-Triangulation (continue tracks with low observation errors and merge them, to improve the completeness of the reconstruction) --> Iterative Refinement (BA->RT->Filtering)

Redundant View Mining

After BA, we partition the unaffected scene parts into groups of highly overlapping images and parameterize each group as a single camera, to achieve a computational benefit.


  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
Bayesian Color Constancy is a computational framework that attempts to solve the problem of color constancy in computer vision. Color constancy refers to the ability of the human visual system to perceive the colors of objects as relatively constant, despite changes in the spectral composition of the illuminating light. The Bayesian Color Constancy framework is based on the idea that the visual system should use prior knowledge about the statistical properties of the colors in the scene, as well as the statistical properties of the illuminant, to infer the true colors of the objects in the scene. This is done by formulating a probabilistic model of the relationship between the colors of the objects and the illuminant, and then using Bayesian inference to estimate the most likely illuminant given the observed colors. The paper "Bayesian Color Constancy Revisited" presents a review of the Bayesian Color Constancy framework, and proposes several improvements to the original model. One of the main contributions of the paper is the development of a new method for estimating the prior distribution of illuminants, based on a statistical analysis of the colors in a large set of natural images. The authors also propose several modifications to the likelihood function used in the model, in order to better capture the properties of real-world images. Overall, the paper provides a comprehensive overview of the Bayesian Color Constancy framework, and presents several important improvements and extensions to the original model. These developments have the potential to improve the accuracy and robustness of color constancy algorithms in computer vision applications.
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值